Improving care for depression is a fundamental goal of mental health services research. The need to improve care within primary care is particularly relevant because most patients receive care solely within that setting. Primary care provider (PCP) behavior is a key vehicle for achieving improvement. This application (responding to PAR99-073) aims to enhance the understanding of underlying relationships among provider attitudes, intentions, and PCP treatment behavior to improve care for depression. The multidisciplinary research team will examine and compare selected theories from the social and behavioral sciences to gain new insights into PCP depression treatment behavior. This work will ask old questions about improving mental health care in new ways by applying and integrating models traditionally used to study personal health behavior to study PCP depression treatment behavior. Lessons from application of these frameworks could inform future interventions for changing provider behavior to improve depression care. These objectives will be addressed through secondary analyses of data from a sample of 414 PCPs and 2,030 patients (5-15 per provider) with major depression participating in the four Quality Improvement for Depression consortium studies. These PCPs are from 80 different clinics in 11 different managed care organizations across the U.S. Few databases contain the breadth of constructs necessary for broad theoretical testing with a relatively large sample of providers. Regression and structural equation analyses will elucidate the degree to which the selected expectancy-value frameworks from social and health psychology, and agency theory from behavioral economics explain provider depression treatment behavior. The relative efficacy of these adapted theories will be compared (using indicators of variance explained and relative goodness of fit) to identify the most useful combination of PCP attitudes, behavioral intentions, and treatment choices in modeling PCP depression treatment behavior. We will also outline the strengths and weaknesses of the models from the different disciplines.